Weight Divergence Driven Divide-and-Conquer Approach for Optimal Federated Learning from non-IID Data.
Pravin ChandranRaghavendra BhatAvinash ChakravarthiSrikanth ChandarPublished in: CoRR (2021)
Keyphrases
- synthetic data
- data collection
- image data
- data sets
- data processing
- prior knowledge
- statistical analysis
- training data
- high quality
- learning process
- data sources
- raw data
- distributed data
- high dimensional data
- data distribution
- online learning
- data mining techniques
- original data
- learning algorithm
- machine learning
- learning tasks
- data objects
- learning models
- bayesian methods
- input data
- missing data
- learning systems
- computer systems
- supervised learning
- knowledge discovery
- active learning
- digital libraries
- data structure
- reinforcement learning
- data mining